Part 1: How to Decide What Data is Worth Collecting

You understand why clean data matters. You know forms are better than spreadsheets. You're ready to start collecting data.

But what data should you collect?

The trap

Businesses want to collect everything. Customer details, transaction history, employee activities, supplier performance, market trends. The list never ends.

So they build forms with twenty fields. They ask staff to log every detail. They imagine dashboards showing every metric.

Three months later, the forms are half-filled. Six months later, they're abandoned. The data is too incomplete to trust. The dashboards show nothing useful.

The system didn't fail because people were lazy. It failed because it asked for too much.

The insight

Data collection only works when it solves a problem people feel right now.

Not a problem that might matter someday. Not a metric that could be interesting. A problem that's causing pain today.

When the sales team doesn't know which leads to prioritise, they'll track lead sources. When the warehouse keeps running out of stock, they'll log inventory movements. When projects keep missing deadlines, they'll record milestones.

Pain creates motivation. Motivation creates consistency. Consistency creates useful data.

Without pain, data collection is just paperwork.

The question to ask

Before building any data collection system, ask: "What problem will this solve that we're struggling with right now?"

If you can't answer specifically, don't build it yet.

"We want to understand our customers better" is not specific enough.

"We don't know why customers leave after three months, and it's killing our revenue" is specific. Now you know what to collect.

Less is more

Here's the counterintuitive part: simple data is often more powerful than complex data.

Stock traders analyse markets using just five numbers per day: Open, High, Low, Close, Volume. From these five fields, they derive hundreds of indicators. Moving averages, volatility measures, trend signals, support levels.

Five fields. Thousands of insights.

They could collect more. News sentiment, weather patterns, social media mentions. But those five fields, collected consistently for decades, are more valuable than a hundred fields collected sporadically.

The lesson: collect the minimum data needed to solve your problem. Collect it without fail. Extract insights through analysis, not by adding more fields.

Start with one problem

Don't design a system to track everything. Design a system to solve one problem.

Get it working. Get the data flowing. Get people into the habit. See results.

Then add another problem. Then another.

This is how sustainable systems are built. Not with grand plans, but with small wins that compound.

What's next?

You know what data to collect. But knowing what to collect doesn't mean people will collect it. Systems fail not because the data wasn't valuable, but because collection never became automatic. In Part 2, we'll cover how to make data collection stick.